An immunochemical comparison of human myelin basic protein and its modified, citrullinated form, C8

Lyso-Lipid-Induced Oligodendrocyte Maturation Underlies Restoration of Optic Nerve Function

Protein hyperdeimination and deficiency of lyso-phospholipids (LPC 18:1) has been associated with the pathology of demyelinating disease in both humans and mice. We uncovered interesting biology of LPC 18:1, in which LPC 18:1 induced optic nerve function restoration through oligodendrocyte maturation and remyelination in mouse model systems. Our in vitro studies show LPC 18:1 protection against neuron-ectopic hyperdeimination and stimulation of oligodendrocyte maturation, while in vivo investigations recorded optic nerve function improvement following optic nerve injections of LPC 18:1, in contrast with LPC 18:0. Thus, just a change in a single bond renders a dramatic alternation in biological function. The incorporation of isobaric C13-histidine in newly synthesized myelin proteins and quantitative proteome shifts are consistent with remyelination underlying restoration in optic nerve function. These results suggest that exogenous LPC 18:1 may provide a therapeutic avenue for stemming vision loss in demyelinating diseases.

Novel Approaches in the Immunotherapy of Multiple Sclerosis: Cyclization of Myelin Epitope Peptides and Conjugation with Mannan

Multiple Sclerosis (MS) is a serious autoimmune disease. The patient in an advanced state of the disease has restrained mobility and remains handicapped. It is therefore understandable that there is a great need for novel drugs and vaccines for the treatment of MS. Herein we summarise two major approaches applied for the treatment of the disease using peptide molecules alone or conjugated with mannan. The first approach focuses on selective myelin epitope peptide or peptide mimetic therapy alone or conjugated with mannan, and the second on immune-therapy by preventing or controlling disease through the release of appropriate cytokines. In both approaches the use of cyclic peptides offers the advantage of increased stability from proteolytic enzymes. In these approaches, the synthesis of myelin epitope peptides conjugated to mannan is of particular interest as this was found to protect mice against experimental autoimmune encephalomyelitis, an animal model of MS, in prophylactic and therapeutic protocols. Protection was peptide-specific and associated with reduced antigen-specific T cell proliferation. The aim of the studies of these peptide epitope analogs is to understand their molecular basis of interactions with human autoimmune T-cell receptor and a MS-associated human leucocyte antigen (HLA)-DR2b. This knowledge will lead the rational design to new beneficial non-peptide mimetic analogs for the treatment of MS. Some issues of the use of nanotechnology will also be addressed as a future trend to tackle the disease. We highlight novel immunomodulation and vaccine-based research against MS based on myelin epitope peptides and strategies developed in our laboratories.

Cyclic citrullinated MBP87-99 peptide stimulates T cell responses: Implications in triggering disease

Amino acid mutations to agonist peptide epitopes of myelin proteins have been used to modulate immune responses and experimental autoimmune encephalomyelitis (EAE, animal model of multiple sclerosis). Such amino acid alteration are termed, altered peptide ligands (APL). We have shown that the agonist myelin basic protein (MBP) 87-99 epitope (MBP_87-99) with crucial T cell receptor (TCR) substitutions at positions 91 and 96 (K⁹¹,P⁹⁶ (TCR contact residues) to R⁹¹,A⁹⁶; [R⁹¹,A⁹⁶]MBP_87-99) results in altered T cell responses and inhibits EAE symptoms. In this study, the role of citrullination of arginines in [R⁹¹,A⁹⁶]MBP_87-99 peptide analog was determined using in vivo experiments in combination with computational studies. The immunogenicity of linear [Cit⁹¹,A⁹⁶,Cit⁹⁷]MBP_87-99 and its cyclic analog - cyclo(87-99)[Cit⁹¹,A⁹⁶,Cit⁹⁷]MBP_87-99 when conjugated to the carrier mannan (polysaccharide) were studied in SJL/J mice. It was found that mannosylated cyclo(87-99)[Cit⁹¹,A⁹⁶,Cit⁹⁷]MBP_87-99 peptide induced strong T cell proliferative responses and IFN-gamma cytokine secretion compared with the linear one. Moreover, the interaction of linear and cyclic peptide analogs with the major histocompatibility complex (MHC II, H2-IA^s) and TCR was analyzed using molecular dynamics simulations at the receptor level, in order to gain a better understanding of the molecular recognition mechanisms that underly the different immunological profiles of citrullinated peptides compared to its agonist native counterpart MBP_87-99 epitope. The results demonstrate that the citrullination of arginine in combination with the backbone conformation of mutated linear and cyclic analogs are significant elements for the immune response triggering the induction of pro-inflammatory cytokines.

Anti-myelin antibodies modulate clinical expression of childhood multiple sclerosis

Anti-myelin basic protein (MBP) antibodies in pediatric-onset MS and controls were characterized. Serum samples were obtained from 94 children with MS and 106 controls. Paired CSF and serum were obtained from 25 children with MS at time of their initial episode of acute demyelinating syndrome (ADS). Complementary assays were applied across samples to evaluate the presence, and the physical binding properties, of anti-MBP antibodies. While the prevalence and titers of serum anti-MBP antibodies against both immature and mature forms of MBP were similar in children with MS and in controls, binding characteristics and formal Surface Plasmon Resonance (SPR) studies indicated surprisingly high binding affinities of all pediatric anti-MBP antibodies. Serum levels of anti-MBP antibodies correlated significantly with their CSF levels, and their presence in children with MS was associated with significantly increased risk of an acute disseminated encephalomyelitis-like initial clinical presentation. While antibodies to both immature and mature forms of MBP can be present as part of the normal pediatric humoral repertoire, these anti-myelin antibodies are of surprisingly high affinity, can access the CNS during inflammation, and have the capacity to modulate disease expression. Our findings identify an immune mechanism that could contribute to the observed heterogeneity in spectrum of clinical presentations in early-onset MS.

gammadelta T cells: the overlooked T-cell subset in demyelinating disease

gammadelta T cells represent a small subpopulation of T cells expressing a restricted repertoire of T-cell receptors and, unlike alphabeta T cells, function more as cells of the innate immune system. These cells are found in skin and mucosal sites as well as secondary lymphoid tissues and frequently act as first line of defense sentinels. gammadelta T cells have been implicated in the pathogenesis of demyelinating disease, although little was known regarding their trafficking and effector functions. In this Mini-Review, we highlight recent studies demonstrating that gammadelta T cells migrate rapidly to the CNS during experimental autoimmune encephalomyelitis (EAE), the animal model for multiple sclerosis. gammadelta T-cell trafficking to the CNS is independent of beta(2)-integrins and occurs well before onset of clinical signs of disease, peaking early during the acute phase of disease. gammadelta T-cell-mediated production of inflammatory cytokines, including interferon-gamma and tumor necrosis factor-alpha, appears critical for EAE development, suggesting that these cells may set the stage for activation of other subsets of infiltrating effector cells. These data suggest that gammadelta T cells or subsets of gammadelta T cells may represent a new therapeutic target in demeylinating disease.

A Tale of Two Citrullines—Structural and Functional Aspects of Myelin Basic Protein Deimination in Health and Disease

Myelin basic protein (MBP) binds to negatively charged lipids on the cytosolic surface of oligodendrocyte membranes and is responsible for adhesion of these surfaces in the multilayered myelin sheath. The pattern of extensive post-translational modifications of MBP is dynamic during normal central nervous system (CNS) development and during myelin degeneration in multiple sclerosis (MS), affecting its interactions with the myelin membranes and with other molecules. In particular, the degree of deimination (or citrullination) of MBP is correlated with the severity of MS, and may represent a primary defect that precedes neurodegeneration due to autoimmune attack. That the degree of MBP deimination is also high in early CNS development indicates that this modification plays major physiological roles in myelin assembly. In this review, we describe the structural and functional consequences of MBP deimination in healthy and diseased myelin.

Immunology of multiple sclerosis

Multiple sclerosis (MS) develops in young adults with a complex predisposing genetic trait and probably requires an inciting environmental insult such as a viral infection to trigger the disease. The activation of CD4+ autoreactive T cells and their differentiation into a Th1 phenotype are a crucial events in the initial steps, and these cells are probably also important players in the long-term evolution of the disease. Damage of the target tissue, the central nervous system, is, however, most likely mediated by other components of the immune system, such as antibodies, complement, CD8+ T cells, and factors produced by innate immune cells. Perturbations in immunomodulatory networks that include Th2 cells, regulatory CD4+ T cells, NK cells, and others may in part be responsible for the relapsing-remitting or chronic progressive nature of the disease. However, an important paradigmatic shift in the study of MS has occurred in the past decade. It is now clear that MS is not just a disease of the immune system, but that factors contributed by the central nervous system are equally important and must be considered in the future.

Expression of citrullinated proteins in murine experimental autoimmune encephalomyelitis

In this study, we demonstrate for the first time the immunohistochemical expression of citrullinated proteins in the central nervous system (CNS) of mice with myelin oligodendrocyte glycoprotein (MOG)-induced experimental autoimmune encephalomyelitis (EAE). By using an established monoclonal antibody (F95) against natural and synthetic citrullinated proteins (Nicholas and Whitaker [2002] Glia 37:328-336), numerous, small, previously unrecognized "patches" of citrullinated proteins were discovered throughout EAE brains, whereas EAE spinal cords showed similar but much larger lesions. On dual color immunofluorescence, these lesions were found to contain citrullinated myelin basic protein (MBP) and were surrounded by astrocytes immunoreactive for both glial fibrillary acidic protein (GFAP) and F95. These lesions became evident about the time when EAE mice became symptomatic and increased in size and number with increasing disease severity. In some sections of spinal cord but not brains of severely debilitated EAE mice, a widespread gliotic response was seen, with astrocytes containing citrullinated GFAP spread throughout the gray and white matter. Western blot analysis of acidic proteins from the brains and spinal cords of EAE mice had higher levels of multiple citrullinated GFAP isoforms compared with controls, with more F95-positive bands in the EAE brains vs. spinal cords. These results raise the possibility that citrullination of both GFAP and MBP may contribute to the pathophysiology of EAE and that the brains of EAE mice may contain more pathology than previously realized.

Immunohistochemical localization of citrullinated proteins in adult rat brain

By using hybridoma technology, an IgM monoclonal antibody (F95) against multiple citrullinated synthetic and natural peptides was recently developed and used to stain immunohistochemically subsets of astrocytes and myelin basic protein (MBP) from selected regions of human brain (Nicholas and Whitaker [2002] Glia 37:328-336). With this antibody, the present study provides a more detailed localization of citrullinated epitopes in the central nervous system (CNS) by examining immunohistochemical staining patterns for F95 in the normal adult rat brain. Thus, immunohistochemical labeling for citrullinated epitopes was seen in white matter areas consistent with myelin staining; however, in general, it was more prominent and uniform in the caudal CNS (spinal cord, medulla oblongata, pons, and cerebellum) than in more rostral areas. F95 staining was also seen in cells and fibers often intimately associated with blood vessels and/or ventricular surfaces. By using dual-color immunofluorescence, the vast majority of this latter staining was colocalized within a subset of astrocytes also immunoreactive for glial fibrillary acidic protein (GFAP). By using Western blot analysis of rat brain proteins, multiple GFAP- and MBP-immunoreactive proteins and peptide fragments were seen, and many of them were also reactive with the F95 antibody. Thus, the present study not only demonstrates that citrullinated epitopes in normal rat brain are most concentrated in subsets of myelin and astrocytes but also provides evidence that GFAP, like MBP, may be present as multiple citrullinated isoforms.

Citrullinated myelin basic protein induces experimental autoimmune encephalomyelitis in Lewis rats through a diverse T cell repertoire

An increased proportion of citrullinated MBP (MBP-C8) occurs in the brains of multiple sclerosis (MS) patients. In this study, MBP-C8 from guinea pig (GP) brains was isolated and found encephalitogenic in Lewis rats upon immunization. An encephalitogenic T cell line selected with MBP-C8 preferentially reacted with MBP-C8 over unmodified MBP. This T cell line responded weakly to the dominant encephalitogenic epitope, GP-MBP peptide 70-88, and did not display restricted TCR beta-chain usage (such as Vbeta88.2). The distinctive features of MBP-C8 were also demonstrated by its ability to reinduce active EAE in 70% of rats which had recovered from unmodified MBP induced EAE. These findings raise the possibility that citrullinated MBP may elicit a different pathogenic T cell repertoire for the recurrent phases of inflammatory demyelination.

Myelin basic protein-like material in the urine of multiple sclerosis patients: relationships to clinical and neuroimaging changes

Urinary myelin basic protein-like material (MBPLM) represents material which is cross-reactive with a cryptic epitope in peptide 84-89 of human myelin basic protein. While normally present at moderate levels in the adult, these levels rise higher in patients who have secondary progressive multiple sclerosis (MS). The increase in urine MBPLM correlates with the burden of disease detected by T2-weighted cranial magnetic resonance imaging. There is no correlation between urinary MBPLM and acute disease activity in relapsing-remitting MS. The first major need for improving the clinical utility of measurements of MBPLM in urine in MS patients is to delineate its exact chemical features so that assays may be improved and a potential biological role of the MBPLM better understood. The second major task is to apply the group data accumulated and apply them to individual patients. This could prove to be means to individually direct treatment and determine its effectiveness.

Myelin basic protein in cerebrospinal fluid and other body fluids

Myelin basic protein (MBP) or a fragment thereof may enter cerebrospinal fluid (CSF) and other body fluids in an etiologically nonspecific fashion to provide information about the status of central nervous system (CNS) myelin damage. MBP immunochemically detected is referred to as MBP-like material (MBPLM). The clinical utility of the assay for MBPLM in CSF is to document the presence, continuation, or resolution of CNS myelin injury. The analysis of CSF for MBPLM is subject to many variables, among which are the antisera and the form of the assay utilized. The dominant epitope of CSF MBPLM is in the decapeptide of 80-89 from the intact MBP molecule of 170 residues. Normally, CSF has no detected MBPLM. Following an acute relapse of MS, MBPLM rises quickly in the range of ng/ml and rapidly declines and disappears. The presence of MBPLM in CSF in chronic and progressive phases of the disease is unusual, but it may sometimes be detected in low levels, depending on the assay used for detection. The level of CSF MBPLM is related to both the mass of CNS myelin damage and how recently it occurred. The level of CSF MBPLM rarely is elevated in optic neuritis. The level of CSF MBPLM is unrelated to CSF protein level, level of IgG, presence of oligoclonal bands or pleocytosis. CSF MBPLM has the potential of serving as a marker of therapeutic effectiveness in MS and does have predictive value for response to glucocorticoids given for worsening of disease. The detection of MBPLM in body fluids other than CSF would be of great value because of the resulting improved feasibility for objectively monitoring the natural history of MS and response to therapy. Studies on blood have yet to produce a valid assay of MBPLM. Urinary MBPLM, though different in its features from that in CSF, may provide a correlate, not with acute demyelination in MS as is the case for CSF, but with progression of disease.

Murine V lambda x and V lambda x-containing antibodies bind human myelin basic protein

Myelin basic protein (MBP) is highly immunogenic and a known autoantigen capable of inducing experimental allergic encephalomyelitis (EAE), the animal model of multiple sclerosis. We have previously described a murine monoclonal antibody (mAb), F28C4, directed against the encephalitogenic MBP peptide acetyl (Ac) 1-9, which contains a V lambda x light chain. Considering the rarity of V lambda x usage, we determined whether other Abs having V lambda x light chains shared similar antigen (Ag) specificity. We screened a panel of V lambda x-containing monoclonal and polyclonal Abs, of unknown specificity for reactivity with MBP. All such Ab, but not heavy chain isotype matched controls, bound MBP but were not polyreactive with other potential self Ags. The binding of a recombinant form of V lambda x alone to MBP demonstrated the important contribution of the V lambda x light chain to the reaction. With the exception of mAb F28C4 which recognizes MBP Ac1-9, the epitope specificity of all other V lambda x-bearing Abs was localized to MBP residues 25-34. These results demonstrate a unique association between V lambda x expression and MBP reactivity. Given that V lambda x shares sequence homology with T cell receptors (TCR) from encephalitogenic T lymphocytes, these results imply a potential role for V lambda x in the pathogenesis of EAE.

The effects of citrullination or variable amino-terminus acylation on the encephalitogenicity of human myelin basic protein in the PL/J mouse

The post-translational modifications of myelin basic protein (MBP) in the form of citrullination and varying length of amino-terminus acylation may modify the biological functions and immunological features of MBP. Both modifications influence the reaction of antibodies and specific T cells recognizing MBP. The present study was undertaken to compare the encephalitogenicity of the citrullinated isomer of MBP (MBP-C8) with the unmodified isomer of MBP (MBP-C1) and to determine if the length of amino-terminal acylation of MBP peptide 1-21 altered an encephalitogenic epitope. MBP-C8, whether from patients with or without multiple sclerosis (MS), and MBP-C1 could induce active experimental allergic encephalomyelitis (EAE) in PL/J mice. A trend of reduced severity of EAE was observed in MBP-C8-injected animals. An increase in the length of amino-terminus fatty acid decreased the encephalitogenicity of MBP peptide 1-21 for both active and adoptive EAE in PL/J mice. Only lymph node cells sensitive to MBP peptide acetyl 1-21 and butyl 1-21 could transfer clinical EAE. In adoptive EAE, MBP peptides hexyl and octyl 1-21 induced moderate histopathological but no clinical change, whereas MBP peptide decyl 1-21 caused neither. A broadening in the antibody response could be detected in the sera of mice with active EAE induced by MBP-acylated peptides 1-21. Our findings demonstrate that encephalitogenicity is retained in the presence of citrullination but that the length of amino-terminus acylation diminishes the encephalitogenicity of MBP in the PL/J mouse.(ABSTRACT TRUNCATED AT 250 WORDS)

Reactivity of normal T-cell lines to MBP isolated from normal and multiple sclerosis white matter

T-cell reactivity to human myelin basic protein (MBP) has been extensively studied using T-cell lines and clones generated from both peripheral blood and cerebrospinal fluid, from normal controls and multiple sclerosis (MS) patients. These studies have largely utilized myelin basic protein isolated from control human adult white matter. In our study, we used MBP reactive T-cell lines as a probe to investigate antigenic differences in a series of MBP preparations isolated from either control human white matter or white matter from the central nervous system (CNS) of MS patients. Autologous peripheral blood derived mononuclear cells were used as antigen presenting cells (APC). Although the majority of T-cells were found to react equally well with all preparations of MBP isolated from both control and MS white matter, we were also able to identify T-cell lines which reacted well with all preparations of MBP isolated from controls but failed to react with MBP isolated from MS white matter. These differences were unlikely to reflect differences in degradation products or excess peptides present in the MS brain since SDS-PAGE and HPLC did not show any difference in the MS samples compared to the controls, and the concentration response curves for a human T-cell clone specific for the 84-102 region of MBP were similar for all the MBP preparations. We did not detect differences in amino acid content amongst MBP preparations although single amino acid substitutions cannot be ruled out. These results raise the possibility that MBP isolated from MS brain may differ in charge microheterogeneity which would affect antigenic determinants.(ABSTRACT TRUNCATED AT 250 WORDS)

Immunological aspects of experimental allergic encephalomyelitis and multiple sclerosis

Multiple sclerosis (MS) is the most frequent, demyelinating disease of the central nervous system (CNS) in Northern Europeans and North Americans. Despite intensive research its etiology is still unknown, but a T cell-mediated autoimmune pathogenesis is likely to be responsible for the demyelination. This hypothesis is based both on findings in MS patients and studies of an experimental animal model for demyelinating diseases, experimental allergic encephalomyelitis (EAE). Experiments in EAE have not only demonstrated which myelin antigens are able to induce the demyelinating process but also have determined the characteristics of encephalitogenic T cells, that is, their fine specificity, major histocompatibility complex (MHC) restriction, lymphokine secretion, activation requirements, and T cell receptor (TCR) usage. Based on these findings, highly specific and efficient immune interventions have been designed in EAE and have raised hopes that similar approaches could modulate the disease process in MS. Although the examination of the myelin-specific T cell response in MS patients has shown parallels to EAE, this remains an area of intensive research because a number of questions remain. This review summarizes the important lessons from EAE, examines recent findings in MS, and discusses current concepts about how the disease process develops and which steps might be taken to modulate it.

Immunological analysis of the amino terminal and the C8 isomer of human myelin basic protein

The citrullination and N-terminus acylation of myelin basic protein (MBP) increases the heterogeneity among the MBP isoforms. The present study was undertaken to further characterize the immune response to the citrullinated form (C8) of MBP as well as to the variably acylated N-terminus of MBP. Six well-characterized murine monoclonal antibodies (mAbs) to human MBP-C8 or MBP peptides (four mAbs to MBP acetyl 1-9, one mAb to MBP 10-19 and one mAb to MBP 80-89), one murine T cell line (PL11) to human MBP peptide acetyl 1-9 and one Lewis rat T cell line (RT-1) to guinea pig (GP) MBP peptide 68-88 were used to assess reactivity with MBP-C1, MBP-C8, and MBP peptides including a series of MBP peptide 1-21 containing 0, 2, 4, 6, 8 or 10 carbon fatty acids. Enzyme-linked immunosorbent assay (ELISA) results revealed that all of the mAbs reacted with human MBP-C1 and MBP-C8 except anti-MBP 10-19 and anti-MBP-C8. The former reacted only with MBP-C1 and the latter only with MBP-C8. The presence and length of acylation of MBP peptide 1-21 modified reactivity. Three mAbs to MBP acetyl 1-9 reacted only with acetyl 1-21, and one mAb anti-MBP actyl 1-9 reacted with all of MBP 1-21 preparations whether acylated or not. mAb anti-MBP-C8 generally reacted better with acylated MBP 1-21 having longer fatty acids. The PL11 T cell line strongly proliferated to human MBP-C1, MBP-C8 and MBP acetyl 1-9, responded, but less well, to MBP 1-21 with longer fatty acids and failed to respond to nonacylated MBP peptide 1-21. The RT-1 cell line responded strongly to GP MBP peptide 68-88, marginally to MBP-C8 and failed to respond to MBP-C1 or any of the other MBP peptides. Specific immune responses to different MBP charge isomers and different N-terminal acylating groups of MBP may play a role in immune-mediated demyelination.

The structure and function of central nervous system myelin

Multiple sclerosis (MS) is characterized by the active degradation of central nervous system myelin, a multilamellar membrane system that insulates nerve axons. MS arises from complex interactions between genetic, immunological, infective, and biochemical mechanisms. Although the circumstances of MS etiology remain hypothetical, one persistent theme involves immune system recognition of myelin-specific antigens derived from myelin basic protein, the most abundant extrinsic myelin membrane protein, and/or another equally suitable myelin protein or lipid. Knowledge of the biochemical and physical-chemical properties of myelin proteins, and lipids, particularly their composition, organization, structure, and accessibility with respect to the compacted myelin multilayers, thus becomes central to understanding how and why these antigens become selected during the development of MS. This article focuses on the current understanding of the molecular basis of MS as it may relate to the protein and lipid components of myelin, which dictate myelin morphology on the basis of protein-lipid and lipid-lipid interactions, and the relationship, if any, between the protein/lipid components and the destruction of myelin in pathological situations.

An antibody specific for component 8 of myelin basic protein from normal brain reacts strongly with component 8 from multiple sclerosis brain

Myelin basic protein (MBP) consists of several components or charge isomers (C-1 through C-8) generated by one or a combination of posttranslational modifications. One of these, C-8, has been shown to contain citrulline (Cit) at defined sites formed by deimination of six arginyl residues. This unusual modification has allowed us to raise antibodies specific for this charge isomer only. To do this, a synthetic peptide, Gly-Cit-Cit-Cit-Cit, was coupled to keyhole limpet hemocyanin and injected into rabbits. The antibodies so generated reacted only with C-8 and not with any of the other charge isomers. A second antibody fraction was raised against the synthetic peptide ACitHGFLPCitHR naturally occurring between residues 24 and 33 of C-8 (all other charge isomers contain R instead of Cit at positions 25 and 31). These antibodies preferred C-8 but reacted with the other charge isomers, to the extent of approximately 25-30% of the reactivity shown with C-8. In studies with C-8 from multiple sclerosis (MS) MBP, much greater reactivity was obtained with these antibodies when compared with their reactivity with C-8 from normal MBP. Because the total number of Cit residues in C-8 from MS and normal MBP is the same, the difference in reactivity may be related to structural factors. The antibodies raised with the tetra-Cit peptide were reacted with three pairs of synthetic peptides: 24ARHGFLPRHR33 and ACitHGFLPCitHR; 120GQRPGFGYGGRAS132 and GQCitPGFGYGGCitAS; and 157GGRDSRSGSPMARR170 and GGCitDSRSGSPMACitR. They reacted only with the Cit-containing peptides in the order 157-170 greater than 120-130 greater than 24-33.(ABSTRACT TRUNCATED AT 250 WORDS)